IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 17, NO.
3, MARCH 1999
Channel Estimation for OFDM Systems with
Transmitter Diversity in Mobile Wireless Channels
Ye (Geoffrey) Li, Senior Member, IEEE, Nambirajan Seshadri, Senior Member, IEEE,
and Sirikiat Ariyavisitakul, Senior Member, IEEE
Abstract— Transmitter diversity is an effective technique to
improve wireless communication performance. In this paper,
we investigate transmitter diversity using space-time coding for
orthogonal frequency division multiplexing (OFDM) systems in
high-speed wireless data applications. We develop channel pa-
rameter estimation approaches, which are crucial for the de-
coding of the space-time codes, and we derive the MSE bounds
of the estimators. The overall receiver performance using such
a transmitter diversity scheme is demonstrated by extensive
computer simulations. For an OFDM system with two transmitter
antennas and two receiver antennas with transmission efficiency
as high as 1.475 bits/s/Hz, the required signal-to-noise ratio is
only about 7 dB for a 1% bit error rate and 9 dB for a 10%
word error rate assuming channels with two-ray, typical urban,
and hilly terrain delay profiles, and a 40-Hz Doppler frequency.
In summary, with the proposed channel estimator, combining
OFDM with transmitter diversity using space-time coding is a
promising technique for highly efficient data transmission over
mobile wireless channels.
I. INTRODUCTION
I N orthogonal frequency division multiplexing (OFDM)
[1]–[8], the entire channel is divided into many narrow
parallel subchannels, thereby increasing the symbol duration
and reducing or eliminating the intersymbol interference (ISI)
caused by the multipath environments. On the other hand,
since the dispersive property of wireless channels causes
frequency selective fading, there is higher error probability
for those subchannels in deep fades. Hence, techniques such
as error correction code and diversity have to be used to
compensate for the frequency selectivity. In this paper, we
investigate transmitter diversity using space-time coding for
OFDM systems.
Transmitter diversity is an effective technique for combating
fading in mobile wireless communications, especially when re-
ceiver diversity is expensive or impractical. Many researchers
have studied transmitter diversity for wireless systems. As
indicated in [9] and the references therein, transmitter di-
versity may be based on linear transforms. The performance
gain of linear transform-based diversity with ideal maximum-
likelihood sequence estimation (MLSE) and an arbitrary num-
ber of transmitter antennas is investigated and compared with
receiver diversity in [10]. Some asymptotic properties of
Manuscript received February 1998; revised August 1998 and November
1998.
Y. Li and S. Ariyavisitakul are with AT&T Labs-Research, Red Bank, NJ
07701-7033 USA.
N. Seshadri is with the Communications Research Department, AT&T
Laboratories, Florham Park, NJ 07932-1004 USA
Publisher Item Identifier S 0733-8716(99)01964-2.
0733–8716/99$10.00  1999 IEEE
461
linear transform-based transmitter diversity are derived in [11].
Transmitter diversity with elaborate selection of the signal
constellations has been proposed in [12]–[15]. Transmitter
diversity combined with Reed–Solomon code has been pro-
posed for clustered OFDM in [2]. More recently, space-time
coding [16] has been developed for high data-rate wireless
communications. Space-time coding is characterized by high
code efficiency and good performance; hence, it is a promising
technique to improve the efficiency and performance of OFDM
systems. In [17], space-time coding with OFDM has been
studied. However, decoding of space-time codes requires
channel state information, which is usually difficult to obtain,
especially for time-variant dispersive fading channels. The
work in [17] assumes ideal channel state information. This
paper focuses on parameter estimation for transmitter diversity
using space-time coding in OFDM systems.
Channel parameter estimation has been successfully used
to improve the performance of OFDM systems. For systems
without cochannel interference, with estimated channel param-
eters [4], [5], [7], [18], coherent demodulation can be used
instead of differential demodulation and can achieve a 3–4
dB signal-to-noise ratio (SNR) gain. Moreover, for systems
with receiver diversity, maximal ratio diversity combining
(MR-DC), which is equivalent to minimum mean-square error
diversity combining (MMSE-DC) in the absence of cochannel
interference, can be achieved directly using the estimated
channel parameters. For systems with cochannel interference,
the coefficients for the MMSE-DC must be calculated from
estimated channel parameters and the instantaneous correlation
of the signals from each receiver [19]. However, there is no
literature on parameter estimation for OFDM systems with
transmitter diversity.
For OFDM systems with transmitter diversity using space-
time coding, two or more different signals are transmitted
from different antennas simultaneously. The received signal
is the superposition of these signals, usually with equal aver-
age power. If the channel parameters corresponding to each
transmitter and receiver antenna pair are estimated by the
approach developed previously in [18], the signals from other
transmitter antenna(s) will become interference. The signal-to-
interference ratio will always be less than 0 dB, and the MSE
of the estimation will be very large. Hence, novel parameter
estimation approaches are desired for transmitter diversity
using space–time coding.
In this paper, we first derive an approach exploiting the
delay profile characteristics of typical mobile wireless chan-
462 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 17, NO. 3, MARCH 1999

the th receiver antenna, which is uncorrelated for different

’s, ’s, or ’s. Note that, for systems with transmitter
diversity, the average SNR ratio at the receiver is defined as

SNR (2)

that is, both and are re-

garded as the desired signals. denotes the ensemble
average.
The decoding of a space-time code [16] uses the Viterbi
decoding algorithm with the metrics
Fig. 1. OFDM system with transmitter diversity using space-time code. (3)

where denotes Euclidean norm, and are

nels. Based on this fundamental idea, we develop a sim-
the received signal vector and the estimated channel parameter
plified approach, which automatically estimates the channel
matrix, respectively, defined as
delay profile and reduces the computation complexity using
this information. As demonstrated by computer simulations,
with estimated channel parameters for decoding of space-time
codes, the OFDM system with two transmitter and two receiver
antennas can transmit data at an efficiency of 1.475 bits/s/Hz
while the required SNR is 7 dB for a 1% bit error rate (BER)
and 9 dB for a 10% word error rate (WER). and is the estimated signal vector, defined as
The rest of this paper is organized as follows. Section II
describes transmitter diversity using space-time coding for
OFDM systems and briefly introduces the statistics of mobile
wireless channels. Section III derives a basic and simplified Hence, accurate estimation of channel parameters is the
approach to parameter estimation for OFDM systems with key to the decoding of the space-time codes. In Section III,
transmitter diversity. Section IV analyzes the performance we describe parameter estimation approaches, which exploit
bounds of the proposed estimation approaches and discusses the delay profile characteristics of mobile wireless channels.
the training block design and identifiability conditions. Finally, In the following, we provide useful representations of these
Section V presents computer simulation results to demonstrate characteristics.
the effectiveness of transmitter diversity for OFDM systems
in various mobile wireless environments. B. Characteristics of Mobile Wireless Channels
The complex baseband representation of the mobile wireless
II. TRANSMITTER DIVERSITY FOR OFDM
channel impulse response can be described by [20]
SYSTEMS IN MOBILE WIRELESS CHANNELS
(4)
A. Transmitter Diversity Using Space-Time Coding
An OFDM system with two-branch transmitter diversity where is the delay of the th path, is the corre-
using a space-time code is shown in Fig. 1. At a transmission sponding complex amplitude, and is the shaping pulse, the
time , a binary data block is frequency response of which is usually a square-root raised-
coded into two different signals, , for cosine Nyquist filter. Hence, the frequency response at time
. Each of these signals forms an OFDM block. The two is
transmitter antennas simultaneously transmit OFDM signals
modulated by for . Hence, discrete Fourier
transform (DFT) of the received signal at each receiver antenna
is the superposition of two distorted transmitted signals, which (5)
can be expressed as
with
(1)

where is the channel frequency response for the

th tone at time , corresponding to the th transmitter Due to the motion of the vehicle, ’s are modeled to be
antenna and the th receiver antenna, whose characteristics wide-sense stationary (WSS), narrowband complex Gaussian
will be discussed in Section II-B; denotes the additive processes, which are independent for different paths. Fur-
complex Gaussian noise, with zero mean and variance , on thermore, ’s for all have the same normalized time
LI et al.: CHANNEL ESTIMATION FOR OFDM SYSTEMS 463

correlation function and different average powers . It has

been demonstrated in [18] that the correlation function of
can be separated into the multiplication of a time-
domain correlation and a frequency-domain correlation, which
makes it possible to exploit the correlation in both domains
separately.
For OFDM systems with proper cyclic extension and sample
timing, it has been shown in [4] and [18] that, with tolerable
leakage, the channel frequency response can be expressed as

(6)

where , ,
is the number of tones of an OFDM block, and are (a)
the block length and tone spacing of the OFDM system,
respectively, and is the sample interval of the system
that relates to by . In (6), ’s, for
, are WSS, narrowband complex Gaussian
processes. The average power of and index ( )
depend on the delay profiles and dispersion of the wireless
channels.
The two-ray [21], typical urban (TU), and hilly terrain (HT)
[20], [22] models are three commonly used delay profiles. The
average power and delay of each ray for TU and HT delay
profiles are shown in Fig. 2(a) and (b), respectively. When
comparing the performance of OFDM systems under different
delay profiles, we assume the same delay spread, defined as

(b)
Fig. 2. (a) The TU delay profile (d = 1:06 s) and (b) the HT profile
(d = 5:04 s).

For the two-ray profile with equal average power on each

ray, the delay spread is ( ), i.e., a half of
the delay difference between the two rays. The delay spreads expressed as1
for TU and HT delay profiles in Fig. 2 are 1.06 and 5.04 s,
respectively.
(7)

III. CHANNEL PARAMETER ESTIMATION

Hence, to obtain , we only need to estimate .
As indicated in Section II, to decode a space-time code From Section II, the received signal at each antenna can be
used for transmitter diversity in OFDM systems, the channel expressed as
parameters must be provided. The difficulty of parameter
estimation here is that, for each receiver antenna, every tone
is associated with multiple channel parameters. Fortunately, (8)
channel parameters for the different tones of each channel
are correlated. With this correlation, we are able to develop
parameter estimation approaches. for and all . If the transmitted signals
’s, for , are known through the use of a training
A. Basic Approach block, the temporal estimation of can be found by
Following (6), the frequency response at the th tone of the 1 The index j for different receiver antennas is omitted from Hij [n; k],
th block corresponding to the th transmitter antenna can be rj [n; k], and wj [n; k].
464 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 17, NO. 3, MARCH 1999

minimizing the following MSE cost function: Then, (11) is equivalent to

(14)

(9) for and .

Equation (14) can be written in matrix form as shown at
the bottom of the page in (15) for . Hence,
Hence, can be determined by can be estimated by

(16)

With the temporal estimation of the channel parameters, a

more accurate estimation can be obtained by exploiting the
time correlation (i.e., the correlation between different blocks)
of the channel parameters. Hence, using the results in [18], a
robust channel estimator for OFDM systems with transmitter
(10) diversity can be constructed, as shown in Fig. 3. We have
discussed a procedure for computing coefficients of the filter
for robust estimator in [18], which is ideal -band-
where and denote the real and imaginary part of
limited, where is the maximum Doppler frequency of the
a complex number, respectively. Direct calculation yields that
channel.
(10) is equivalent to
In the above derivation, we have assumed that the transmit-
ted signals for each antenna are known. When the system is
in data transmission mode, decoded data are used to generate
the reference signals.

(11)
B. Simplified Approach
To get the temporal estimation of the channel parameters
for and , where denotes the
using the basic approach developed in Section III-A, has
complex conjugate of the complex number .
to be calculated. For a system with a 160- s symbol duration
Define
and 128 tones, is 17 if the channel delay span is 20 s.
is a 34 34 matrix. Thus, calculating requires intensive
computations. Using the characteristics of the channel delay
(12) profiles, the estimation approach can be greatly simplified.
Most mobile wireless channels are characterized by discrete
multipath arrivals, i.e., the magnitude of ’s for most
and ’s are zero or very small; hence, these channel taps can
be ignored. Using this characteristic, we are able to reduce
the computation complexity and improve the performance of
(13) the parameter estimation by identifying significant taps during
training blocks.

where
and
with

and

.. .. .. .. (15)
. . . .
LI et al.: CHANNEL ESTIMATION FOR OFDM SYSTEMS 465

Fig. 3. Channel parameter estimator for transmitter diversity of OFDM systems.

For packet data transmission in wireless systems, the first As before, the time correlation of can be used
OFDM block of a packet is the training block that is used to further improve its estimation. Let be the time-
for initial channel parameter estimation and for time and averaged estimation; then the channel frequency response is
frequency synchronizations. Since the transmitted data for determined by
the training block is known, in (16) can be calculated
in advance and initial parameter estimation only requires
multiplying by . (17)
Let ’s for and be
the estimated parameters from the training block. Then, we
can identify the significant taps by finding the ’s with large It should be indicated that the simplified approach is with
. Assuming ’s for and zero error only if the nonsignificant taps are actually ze-
( ) to be the ros. Consequently, the choice of depends on both the
significant taps, then we only need to estimate computational complexity and the required performance of
corresponding to these taps and ignore the rest as shown in the simplified estimator. For systems with large SNR, the
estimation error of is small and the estimation error
(16a) at the bottom of the page.
of comes mainly from the leakage; hence, should
Similar to the derivation of Section III-A, the significant
be larger to reduce the leakage. However, for systems with
channel taps for each channel corresponding to the same
small SNR, the relative estimation error for those taps with
receiver antenna and different transmitter antennas can be
small amplitude, is very large; hence, should be smaller.
estimated by

IV. PERFORMANCE ANALYSIS

We have developed a basic and simplified approach in
Here, is a matrix. As demonstrated by Section III. In this section, we analyze the performance and
simulation, is enough for all three delay profiles derive the MSE bounds. We also briefly discuss channel
introduced in Section II-B. identifiability and training signal design.

Let

.. .. .. ..
. . . .

and
(16a)
466 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 17, NO. 3, MARCH 1999

A. MSE Bound of Estimator Similarly

From (8) and (12), we have

where

for . Therefore,

(21)

where

with

(18)
Hence, from (16)
By means of the convolution theorem [23]
(22)

From (22), the (temporal) estimation of is the sum-

mation of the true value plus a term affected by channel
noise.
and
From (22)

(23)

which implies that (16) is an unbiased estimation of .

The MSE of the temporal estimator is

Let MSE

(19)

for . Then, for all Tr

Tr (24)

where denotes the Hermitian of a matrix or vector and Tr

denotes the trace of a square matrix.
If ’s, for , all and , are constant modulus,
then

thus
(20)

Hence
where is a identity matrix. Since ,
the result is shown in (24a) at the bottom of the page.

(24a)
LI et al.: CHANNEL ESTIMATION FOR OFDM SYSTEMS 467

Hence, the average estimation error 3) Enhanced Parameter Estimation: As indicated in [19],
the second-pass channel estimation can also be used here
MSE Tr to improve the performance of the channel estimator. In
that case, the lower bound of the MSE is reduced to

MSE (31)

Tr VI. PERFORMANCE EVALUTAION BY COMPUTER SIMULATION

Extensive computer simulations have been conducted to
demonstrate the performance of transmitter diversity using
(25)
space-time coding. Before presenting the simulation results,
we first describe the parameters of the simulated OFDM
Let ’s, for , be the eigenvalues of systems.
. Then, the MSE can also be expressed as
A. Parameters of OFDM with Transmitter Diversity
MSE (26) In our simulation, channels with two-ray, TU, and HT delay
profiles and a Doppler frequency of 40 and 200 Hz are used
From the above expression, the MSE of the temporal estima- to represent different mobile environments. Two transmitter
tion depends on the transmitted sequence, and furthermore antennas and two receiver antennas are used for diversity. The
links between different transmitter or receiver antennas are
MSE (27) independent; however, they have the same global statistics.
The parameters of the simulated OFDM system are similar
with equality if and only if for , or to those in [3], [18], and [19]. The entire channel band-
equivalently, for , or width, 800 kHz, is divided into 128 subchannels. The four
for all . subchannels on each end are used as guard tones, and the
From [18], the lower bound of the MSE of the estimated rest (120 tones) are used to transmit data. To make the tones
is orthogonal to each other, the symbol duration is 160 s. An
additional 40 s guard interval is used to provide protection
MSE (28) from intersymbol interference due to channel multipath delay
spread. This results in a total block length 200 s and
a subchannel symbol rate 5 kBd.
where is the Doppler frequency of the channel. A 16-state space-time code with four PSK is used in the
system. Each data block, containing 236 bits, is coded into
V. REMARKS two different blocks, each of which has exactly 120 symbols
1) Identification Condition: From (16), the necessary and to form an OFDM block. Consequently, it is in fact a space-
sufficient condition for the channel parameters to be frequency code. The channel parameter estimation approaches
identifiable is that is invertible. From (25), this developed in this paper are used to provide estimated param-
condition is equivalent to that eters for decoding.
being invertible. A sufficient identification condition is The described system can transmit data at a rate of 1.18
that Mbits/s over an 800 kHz channel, i.e., the transmission effi-
ciency is 1.475 bits/s/Hz.
Tr

or B. Simulation Results
The performance of the systems is measured by WER, BER,
(29) and the estimator’s MSE, each averaged over 10 000 OFDM
blocks. Fig. 4 shows the WER, BER, and the estimator’s MSE
of the OFDM system with transmitter diversity for the TU and
2) Training Signal Design: From (27), in order for the two-ray delay profile channels with 40-Hz Doppler frequency
channel estimator to achieve its best performance, the ( ) and the same delay spread 1.06 s. When ideal
training signals from the two transmitter antennas should channel parameters of the previous OFDM block are used
satisfy for decoding of the current OFDM block, the OFDM system
(30) for both delay profiles has almost the same performance, and
the required SNR is 8 and 11 dB for 10 and 1% WER,
for , which can be respectively, and 6 and 8 dB for 1 and 0.1% BER, respectively.
easily constructed, for example, by letting If a nine-tap significant-tap-catching (STC) estimator is used
. to estimate the channel parameters, then there is about a
468 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 17, NO. 3, MARCH 1999

(a) (a)

(b) (b)
Fig. 5. (a) WER and (b) BER of OFDM systems with transmitter diversity
for channels with the two-ray and the TU delay profiles, respectively, and
fd = 200 Hz when different number of taps for the estimators is used.

(c)
Fig. 4. (a) WER, (b) BER, and (c) MSE of OFDM systems with transmitter
diversity for channels with the two-ray and the TU delay profiles, respectively,
and fd = 40 Hz when a different number of taps for the estimators is used.
1-dB degradation in the required SNR’s. However, larger
performance degradation can be seen for a seven or five-
tap STC estimator. Since the estimator has more leakage for
the TU delay profile, the system has a larger performance
degradation than for the two-ray delay profile. Fig. 5 shows
similar performance for hz. Again, the receiver
performance is generally better for the two-ray delay profile
than for the TU delay profile.
Fig. 6 compares the performance between channels with the
two-ray and the HT delay profiles with 40 Hz and
5.04 s. Similar to the above results, for channels with
40 Hz, the system has the same performance when the ideal
parameters of the previous OFDM block are used for decoding.
However, when estimated parameters are used, the system has
better performance for the two-ray delay profile than for the
HT profile, since the estimator has lower MSE for the two-
ray delay profile. When a seven or nine-tap STC estimator is
used, the required SNR is 8 dB for a 10% WER, 6 dB for a
1% BER (for the two-ray delay profile), and about 8.6 dB for
a 10% WER and 6.6 dB for a 1% BER, respectively.
LI et al.: CHANNEL ESTIMATION FOR OFDM SYSTEMS
(a)
(b)
(c)
Fig. 6. (a) WER, (b) BER, and (c) MSE of OFDM systems with transmitter
diversity for channels with the two-ray and the HT delay profiles, respectively,
and fd = 40 Hz when different number of taps for the estimators is used.
As indicated in Section IV-B, enhanced channel estimation
can improve the OFDM system performance. Fig. 7 shows
469
(a)
(b)
Fig. 7. WER’s of original and enhanced estimators (seven-tap STC) for
channels with (a) the two-ray and the TU delay profiles and (b) the two-ray
and the HT delay profiles and with different Doppler frequencies.
the performance improvement of the enhanced estimator for
channels with 40 and 200 Hz, respectively. From the
figure, for channels with a lower Doppler frequency (
40 Hz), the enhanced estimator does not improve the system
performance; but for channels with a higher Doppler frequency
( 200 Hz), the enhanced estimator improves the required
SNR for a 10% WER by over 1 dB.
VII. DISCUSSIONS
In Section VI, we have shown the performance improve-
ment of an OFDM system with transmitter diversity using
a 16-state space-time code. A higher state, such as 64-state,
space-time code can be used to further improve the system
performance; however, the decoding complexity will increase.
It has been shown in [18] that an OFDM system with one
transmitter antenna and two receiver antennas transmitting data
at an efficiency of 0.75 bit/s/Hz requires about 6.5-dB SNR for
a 10% WER. For the simulated system simulated in Section VI
470 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 17, NO. 3, MARCH 1999
with two transmitter antennas, the data transmission rate is
doubled and the required SNR is about 9 dB. Furthermore,
considering the transmitted data rate and the length of codes
(236 bits for the space-time code in this paper and 120 bits for
the RS code in [18]), the required for a 10% WER is
about 1.5 dB less for the OFDM system studied in this paper,
which is the gain of transmitter diversity using the space-time
code.
However, the use of transmitter diversity using space-
time codes may affect the system’s interference suppression
performance since each cochannel interference source would
produce two interference signals for systems with two trans-
mitter antennas. This effect is under investigation.
VIII. CONCLUSIONS
In this paper, we study transmitter diversity using space-time
coding for OFDM systems. We develop channel parameter
estimation approaches, which are crucial for the decoding of
space-time codes, and we derive the MSE bounds for these
estimation approaches. The overall receiver performance is
evaluated by computer simulation. For an OFDM system with
two transmitter antennas and two receiver antennas using
space-time coding, permitting 1.475 bits/s/Hz, the required
SNR is about 9 dB for 10% WER, and 7 dB for 1% BER,
for channels with the two-ray, TU, and HT delay profiles and
a Doppler frequency of 40 Hz. Therefore, OFDM systems
with transmitter diversity using space-time coding can be used
for highly efficient data transmission over mobile wireless
channels.
ACKNOWLEDGMENT
The authors wish to thank V. Tarokh for providing the
space-time coding program. The authors would also like
to thank L. J. Cimini, N. R. Sollenberger, W. Turin, and
the anonymous reviewers for their insightful comments that
significantly improved the quality of the paper.
REFERENCES
[1] L. J. Cimini, Jr., “Analysis and simulation of a digital mobile channel us-
ing orthogonal frequency division multiplexing,” IEEE Trans. Commun.,
vol. COM-33, pp. 665–675, July 1985.
[2] L. J. Cimini, Jr., B. Daneshrad, and N. R. Sollenberger, “Clustered
OFDM with transmitter diversity and coding,” in Proc. 1996 IEEE
Global Telecommunications Conf., London, U.K., pp. 703–707.
[3] L. J. Cimini, Jr. and N. R. Sollenberger, “OFDM with diversity
and coding for high-bit-rate mobile data applications,” in Proc. 1997
IEEE Global Telecommunications Conf., Phoenix, AZ, Nov. 1997, pp.
305–309.
[4] J.-J. van de Beek, O. Edfors, M. Sandell, S. K. Wilson, and P. O.
Börjesson, “On channel estimation in OFDM systems,” in Proc. 45th
IEEE Vehicular Technology Conf., Chicago, IL, July 1995, pp. 815–819.
[5] O. Edfors, M. Sandell, J.-J. van de Beek, S. K. Wilson, and P. O.
Börjesson, “OFDM channel estimation by singular value decomposi-
tion,” IEEE Trans. Commun., vol. 46, pp. 931–939, July 1998.
[6] I. Kalet, “The multitone channel,” IEEE Trans. Commun., vol. 37, pp.
119–124, Feb. 1989.
[7] V. Mignone and A. Morello, “CD3-OFDM: A novel demodulation
scheme for fixed and mobile receivers,” IEEE Trans. Commun., vol.
44, pp. 1144–1151, Sept. 1996.
[8] S. B. Weinstein and P. M. Ebert, “Data transmission by frequency-
division multiplexing using the discrete Fourier transform,” IEEE Trans.
Commun. Technol., vol. 19, pp. 628–634, Oct. 1971.
[9] A. Wittneben, “A new bandwidth efficient transmit antenna modulation
diversity scheme for linear digital modulation,” in Proc. 1993 IEEE Int.
Communications Conf., Geneva, Switerland, June 1993, pp. 1630–1634.
[10] J. H. Winters, “The diversity gain of transmit diversity in wireless
systems with Rayleigh fading,” in Proc. 1994 IEEE Int. Communications
Conf., Chicago, IL, June 1994, pp. 1121–1125.
[11] G. W. Wornell and M. D. Trott, “Efficient signal processing techniques
for exploiting transmit antenna diversity on fading channels,” IEEE
Trans. Signal Processing, vol. 45, pp. 191–205, Jan. 1997.
[12] K. J. Kerpez, “Constellations for good diversity performance,” IEEE
Trans. Commun., vol. 41, pp. 1412–1421, Sept. 1993.
[13] D. Rainish, “Diversity transform for fading channels,” IEEE Trans.
Commun., vol. 44, pp. 1653–1661, Dec. 1996.
[14] V. M. DaSilva and E. S. Sousa, “Fading-resistant modulation using
several transmitter antennas,” IEEE Trans. Commun., vol. 45, pp.
1236–1244, Oct. 1997.
[15] N. Seshadri and J. H. Winters, “Two signaling schemes for improving
the error performance of frequency division duplex (FDD) transmission
systems using transmitter antenna diversity,” Int. J. Wireless Inform.
Networks, vol. 1, no. 1, 1994.
[16] V. Tarokh, N. Seshadri, and A. R. Calderbank, “Space-time codes for
high data rate wireless communication: Performance analysis and code
construction,” IEEE Trans. Inform. Theory, vol. 44, pp. 744–765, Mar.
1998.
[17] D. Agarwal, V. Tarokh, A. Naguib, and N. Seshadri, “Space-time
coded OFDM for high data rate wireless communication over wideband
channels,” in Proc. 48th IEEE Vehicular Technology Conf., Ottawa,
Canada, May 1998, pp. 2232–2236.
[18] Y. Li, L. J. Cimini, Jr., and N. R. Sollenberger, “Robust channel
estimation for OFDM systems with rapid dispersive fading channels,”
IEEE Trans. Commun., vol. 46, pp. 902–915, July 1998.
[19] Y. Li and N. R. Sollenberger, “Adaptive antenna arrays for OFDM
systems with cochannel interference,” IEEE Trans. Commun., vol. 47,
pp. 217–229, Feb. 1999.
[20] R. Steele, Mobile Radio Communications. New York: IEEE Press,
1992.
[21] Y. Li, J. H. Winters, and N. R. Sollenberger, “Spatial-temporal equal-
ization for IS-136 TDMA systems with rapid dispersive fading and
co-channel interference,” IEEE Trans. Veh. Technol., to be published.
[22] S. Ariyavisitakul, N. R. Sollenberger, and L. J. Greenstein, “Tap-
selectable decision feedback equalization,” IEEE Trans. Commun., vol.
45, pp. 1521–1526, Dec. 1997.
[23] A. V. Oppenheim and R. W. Schafer, Discrete-Time Signal Processing.
Englewood Cliffs, NJ: Prentice-Hall, 1989.
Ye (Geoffrey) Li (S’92–M’95–SM’97) was born in
Jiangsu, China. He received the B.S.E. and M.S.E.
degrees in 1983 and 1986, respectively, from the
Department of Wireless Engineering, Nanjing Insti-
tute of Technology, Nanjing, China, and the Ph.D.
degree in 1994 from the Department of Electrical
Engineering, Auburn University, Auburn, AL.
From 1986 to 1991, he was a Teaching Assistant
and then a Lecturer with Southeast University, Nan-
jing, China. From 1991 to 1994, he was a Research
and Teaching Assistant with Auburn University.
From 1994 to 1996, he was a Postdoctoral Research Associate with the
University of Maryland, College Park. Since 1996, he has been with AT&T
Labs-Research, Red Bank, NJ. His general research interests include statistical
signal processing and wireless mobile systems, with emphasis on signal
processing in communications.
Dr. Li is currently serving as a Guest Editor for a Special Issue on Signal
Processing for Wireless Communications for the IEEE JOURNAL ON SELECTED
AREAS IN COMMUNICATIONS and as an Editor for wireless communication
theory for the IEEE TRANSACTIONS ON COMMUNICATIONS.
LI et al.: CHANNEL ESTIMATION FOR OFDM SYSTEMS
Nambirajan Seshadri (S’81–M’82–SM’95) re-
ceived the Bachelor’s degree in electronics and
communications engineering from the University
of Madras, India, in 1982 and the M.S. and Ph.D.
degrees in electrical and computer engineering from
Rensselaer Polytechnic Institute, Troy, NY, in 1984
and 1986, respectively.
He is Head of the Communications Research
Department at AT&T Laboratories, Florham Park,
NJ. Prior to this, he was a Distinguished Member of
Technical Staff at AT&T Bell Laboratories, Murray
Hill, NJ. His research activities in the department cover a broad range of
signal processing and communications concepts, including signal analysis for
compression and transmission, error resilient signal compression techniques,
new transmission techniques for wireless such as space-time coding, radio
link adaptation algorithms and protocol design, and design of a flexible
wireless simulator. He has authored more than 60 conference and journal
papers and holds 20 patents.
Dr. Seshadri was the Associate Editor of coding techniques for IEEE
TRANSACTIONS ON INFORMATION THEORY from 1996 to 1998.
471
Sirikiat Ariyavisitakul (S’85–M’88–SM’93) received the B.S., M.S., and
Ph.D. degrees in electrical engineering from Kyoto University, Kyoto, Japan,
in 1983, 1985, and 1988, respectively.
From 1988 to 1994, he was a Member of the Technical Staff at Bellcore,
Red Bank, NJ, conducting research on wireless communications systems and
experimental prototyping of low-power radio links for personal communi-
cations. He joined AT&T Bell Laboratories in 1994 and continued to work
in the area of wireless communications. He currently works in the Wireless
Communications Research Department of AT&T Laboratories, Red Bank, NJ.
The topics of his research have included modulation techniques, equalization,
linear power amplifiers, cellular CDMA and power control techniques, coding
and frequency hopping, and wireless system architectures and infrastructures.
He holds ten U.S. patents.
Dr. Ariyavisitakul received the 1988 Niwa Memorial Award in Tokyo,
Japan, for outstanding research and publication. He is a member of the Institute
of Electronics, Information and Communication Engineers of Japan. He serves
as the Editor for wireless techniques and fading for IEEE TRANSACTIONS ON
COMMUNICATIONS.